A new hybrid quantitative structure property relationships‐support vector regression (QSPR‐SVR) approach for predicting the solubility of drug compounds in supercritical carbon dioxide

Abstract

The purpose of this work was to compare the performance of 7 meta‐heuristics algorithms namely: Dragonfly (DA), Ant Lion (ALO), Grey Wolf (GWO), Artificial Bee Colony (ABC), Particle Swarm (PSO), Whale (WAO), and a hybrid Particle Swarm with Grey Wolf (HPSOGWO) optimizers in terms of fine‐tuning hyper‐parameters of a hybrid quantitative structure property relationships (QSPR)‐support vector regression (SVR) for the prediction of molar fraction solubilities of drug compounds in supercritical carbon dioxide (SC‐CO2). A dataset of 168 drug compounds, 13 inputs, and 4490 experimental data points was used to achieve the goal. All 7 models were statistically and graphically approved while the HPSOGWO‐SVR was found to over‐perform with an average absolute relative deviation (AARD) of 0.706% and an AIC of −14,434,249. The model was subjected to an external test (validation) using 160 experimental data points that were not used in the training and the test set. The overall results proved that the obtained model has good predictivity ability and robustness.

Hybrid Improved Grey Wolf Support Vector Regression Algorithm for Modeling Solubilities of APIs in Pure Ionic Liquids: σ-Profile Descriptors

“…A model is determined as the best model that represented the experimental result if that model shows the smallest value of AIC (or AICs, BIC) , AIC = N 0.25em ln ( SSE N ) AICs = AICs + 2 n p + 2 n p ( n normalp + 1 ) N − ( n normalp + 1 ) SSE = prefix∑ i = 0 N false( x i obs − x i pred false) 2 BIC = N · ln ( MSE ) + n p · ln ( N ) where N is the number of data points in the data set, x i obs is the observed value, while x̅ obs is its average, x i pred is the predicted value, while x̅ pred is its average, N p is the number of independent variables in each model, MS R is the mean square regression, MS E is the mean square residual, s is the pooled standard deviation, and SSE is the error sum of squares. More details can be found elsewhere. ,− , Typically, a single statistic value is meaningless; the principal model studied in this paper was compared with 11 different topologies.…”

An Arithmetic Optimization Algorithm–Support Vector Regression Approach for Predicting Drug Solubility in Supercritical Carbon Dioxide

“…Although the wide availability of contributions related to this property where some models have made some achievements, the relevant modeling research is still relatively insufficient . The main objective of this study was to apply a support vector machine regression algorithm optimized with an effective meta-heuristic algorithm named Arithmetic Optimization Algorithm (AOA) to build and validate a model able to predict and correlate the solubility of pharmaceutical compounds in supercritical carbon dioxide on a dataset obtained in a previous study [11]. The obtained model AOA-SVR will be examined statistically and graphically.…”